West Nile virus (WN), an agent of significant human and veterinary disease, is endemic in the Old World and rapidly spreading throughout the Americas. Vaccines are needed to halt the geographic expansion of this virus and prevent disease where it is established. However, to preclude introduction of a vaccine virus into the environment, a live attenuated WN vaccine should have low potential for transmission by mosquitoes. A chimeric WN vaccine candidate was previously generated by replacing the membrane and envelope structural protein genes of recombinant dengue type 4 virus (rDEN4) with those of WN; a derivative of this virus, WN/DEN4-3'delta30, contains a 30-nucleotide deletion in the 3' untranslated region. To assess the potential for transmission by mosquitoes of these vaccine candidates, the ability of each chimeric virus to infect the mosquito midgut, disseminate to the head, and pass into the saliva was compared to that of their wild-type parental WN and DEN4 viruses in three vector species. The WN/DEN4 chimeric viruses were significantly attenuated in both Culex tarsalis, a vector able to transmit WN but not dengue, and in Ae. aegypti, a vector able to transmit dengue but not WN. However, the chimeric viruses were as infectious as either wild-type virus for Ae. albopictus, a vector able to transmit both dengue and WN. These results indicate that chimerization caused a contraction in vector host range rather than universal attenuation for mosquitoes per se. This restriction in potential vectors renders it less likely that WN/DEN4 and WN/DEN4-3'delta30 would be transmitted from vaccinees to mosquitoes.